Objects generally emit infrared radiation across a spectrum of wavelengths, but sometimes only a limited region of the spectrum is of interest because sensors usually collect radiation only within a specific bandwidth. Multi-spectral imaging captures image data at frequencies across an electromagnetic spectrum, and can allow extraction of additional information that a human eye may fail to capture with visible red, green and blue light receptors. Multi-spectral imaging divides a spectrum into many bands, in contrast to only recording a total intensity of radiation falling on each pixel.
Wavelengths of frequencies may be separated by filters or by use of instruments that are sensitive to particular wavelengths. For example, light from frequencies may be beyond a visible light range such as infrared light. It may be desirable to acquire one digital image in a band of visible spectra, for example a range of about 0.7 μm to about 0.4 μm, called a red-green-blue (RGB) region, and also in infra-red wavelengths of about 0.7 μm to about 10 μm or more.
A boundary between visible and infrared light is not precisely defined. A human eye is markedly less sensitive to light above about 700 nm wavelength, so longer wavelengths may make limited contributions to scenes illuminated by common light sources. An onset of infrared may be defined according to a variety of standards at various values generally between about 700 nm and about 800 nm wavelength.
Infrared spectrum is often subdivided into smaller sections due to different sensor responses to infrared light. For example, near infrared may be a spectrum region closest in wavelength to radiation detectable by a human eye, and mid infrared and far infrared may be progressively further from a visible spectrum. The infrared spectrum may be divided into bands based on availability of light sources transmitting, absorbing materials (e.g., optical fibers) and detectors. For example, short-wavelength infrared (SWIR), Near InfraRed (NIR), Middle InfraRed (MIR), and Far InfraRed (FIR) or thermal, are commonly used references for frequency bands. SWIR may generally be variously defined by, for example, about 1 to about 3 μm, about 1.4 to about 3 μm, or similar range. Spectral imaging with numerous bands, fine spectral resolution, and wide spectral coverage may be called “hyperspectral”.